Prognostic value of esophageal cancer immune prognostic index in advanced esophageal squamous cell carcinoma patients with anti‐programmed cell death‐1 therapy

Abstract Background This study aimed to determine whether the immune prognostic index (ECIPI), based on hemoglobin (Hb) and neutrophil‐to‐lymphocyte ratio (NLR), could predict the prognosis in patients with advanced esophageal squamous cell carcinoma (ESCC) receiving programmed cell death‐1 (PD‐1) inhibitor treatment. Methods Advanced ESCC patients who had been treated with PD‐1 inhibitors from Jan 2016 to Oct 2021 were included. Kaplan–Meier method and Cox proportional hazards regression were used to analyze progression‐free survival (PFS) and overall survival (OS). The overall response rate (ORR) was the percentage of complete and partial responses. Univariate and multivariate analyses were used for estimating hazard ratio (HR) and 95% confidence interval (CI). Patients were grouped by ECIPI (good: Hb > 105 g/L and NLR ≤ 4.3; intermediate: Hb ≤ 105 g/L and NLR ≤ 4.3, or Hb > 105 g/L and NLR < 4.3; poor: Hb ≤ 105 g/L and NLR > 4.3). Variables for the multivariate model were selected if the p‐value was below 0.05 in the univariate analysis. All statistical comparisons were two‐way, and a p‐value below 0.05 was set as statistical significance. Results Totally, of 123 ESCC patients with stage III or IV were included in the study. Efficacy evaluation showed that patients with pretreatment ECIPI good had the best ORR compared with those with ECIPI intermediate and ECIPI poor (53% vs. 22% vs. 8%, p < 0.01). Multivariate analysis showed that ECIPI was an independent influential factor for PFS (p = 0.004) and OS (p < 0.001). Kaplan–Meier curves demonstrated that patients with ECIPI good had the longest PFS (median: 11.6 vs. 3.5 vs. 1.7 months, p < 0.0001) and OS (median: 23.6 vs. 16.7 vs. 4.0 months, p < 0.0001) compared with those with ECIPI intermediate and ECIPI poor. Subgroup analysis indicated that ECIPI good was associated with improved PFS and OS in patients with ECOG 0–1, PD‐1 inhibitor plus chemotherapy, first‐line treatment, and smoke (all p < 0.05). Conclusions Pretreatment ECIPI was associated with the prognosis in advanced ESCC patients with anti‐PD‐1 therapy, suggesting that ECIPI may be a useful tool to identify patients likely sensitive to PD‐1 inhibitors.


| INTRODUCTION
Esophageal cancer (EC) ranks as the seventh most common cancer, and the sixth leading cause of cancer-related death threat seriously to human health worldwide. 1 Esophageal squamous cell carcinoma (ESCC) is the main histological subtype, which accounts for about 90% of EC. 2,3 For most ESCCs are metastatic at diagnosis, systemic chemotherapy remains the main treatment option with a 5-year survival rate below 5%. 4 Thus, new drugs and treatment strategies need to be developed to improve the therapeutic effect.
In recent years, immunotherapy has been recognized as an exciting therapeutic strategy in various cancers. 5 By inhibiting the pathway of immune checkpoints, the immune system of patients can be activated to fight against tumor cells. 6 Particularly, inhibition of programmed death-1 (PD-1) and programmed death-ligand 1 (PD-L1) have gained enormous clinical utility in a variety of malignancies including ESCC. 7,8 However, these benefits are limited to a small percentage of patients. Therefore, there is an urgent need to identify patients likely to respond to immunotherapy.
Anemia is a common hematologic abnormality in advanced cancer patients. 9 Previous research showed that pretreatment hemoglobin (Hb) level was associated with the prognosis of non-small cell lung cancer (NSCLC) patients receiving immunotherapy. 10 Meanwhile, growing research has reported the key role of inflammation in the process of cancer initiation, development, and metastasis. 11 The neutrophil-to-lymphocyte ratio (NLR), calculated by absolute neutrophil count (ANC) and absolute lymphocyte count (ALC), represents the inflammatory state of tumors and has been investigated in a variety of solid tumors including EC. [12][13][14] However, no studies have yet described the predictive value of esophageal cancer immunological prognostic index (ECIPI, combining Hb and NLR) in advanced ESCC patients receiving immunotherapy. Therefore, this study aimed to investigate the association between pretreatment ECIPI and prognosis in advanced ESCC patients treated with PD-1 inhibitors.

| Patients
Patients with ESCC who had been treated with PD-1 inhibitors from Jan 2016 to Oct 2021 were identified in the Chinese PLA General Hospital. Patients should be excluded according to the following criteria: (I) patients received one cycle of PD-1 inhibitor treatment; (II) baseline Hb, ANC, and ALC were not measured before initial immunotherapy (within 5 days); (III) treatment efficacy was not evaluated after initial immunotherapy (generally 6 weeks); (IV) patients with TNM stage I or II.

| Data extraction and definition
Two independent investigators (J.Y. Lu and Z.B. Zhang) extracted data from medical records, including age, gender, clinical stage, smoking history, Eastern Cooperative Oncology Group performance status (ECOG PS), treatment strategy (PD-1 inhibitor plus chemotherapy or radiotherapy), treatment line of immunotherapy (1-line, 2-line, and ≥ 3-line), treatment response evaluation, blood results of Hb, ANC, and ALC at baseline. Disagreements were resolved through discussion with the other two investigators (X. Lei and L.H. Du).
The ECIPI was developed based on Hb and NLR, and the cutoff values of Hb and NLR were estimated by X-tile v3.6.1. 15 Patients were divided into three groups by ECIPI (good: Hb > 105 g/L and NLR ≤ 4.3; intermediate: Hb ≤ 105 g/L and NLR ≤ 4.3, or Hb > 105 g/L and NLR < 4.3; poor: Hb ≤ 105 g/L and NLR > 4.3). The Response Evaluation Criteria in Solid Tumors (v1.1) was used for evaluating treatment efficacy. 16 Overall response rate (ORR) referred to the rate of complete response (CR) and partial response (PR). Progression-free survival (PFS) was the interval from initiation of immunotherapy to disease progression, death, or the last follow-up (if censored). Overall survival (OS) was the interval from the initiation of immunotherapy to death or the last follow-up (if censored). All patients had been followed up by telephone and searching medical records until Jan 29, 2022.
Conclusions: Pretreatment ECIPI was associated with the prognosis in advanced ESCC patients with anti-PD-1 therapy, suggesting that ECIPI may be a useful tool to identify patients likely sensitive to PD-1 inhibitors.

K E Y W O R D S
efficacy, esophageal cancer immune prognostic index, esophageal squamous cell carcinoma, prognosis, programmed cell death-1 inhibitor

| Statistical analysis
All statistical analyses were performed using SPSS v23.0 and graphs were drawn using GraphPad Prism 8.0. The cutoff values for Hb and NLR were determined with X-tile v3.6.1. Comparations between categorical variables were performed using Pearson's chi-square test or Fisher's exact test. The Kaplan-Meier method was used for survival analysis and the log-rank test for comparison. Cox proportional hazards regression models were used for identifying independently influential factors and estimating their hazard ratios (HR) and 95% confidence intervals (CI). Variables included in the multivariate model were selected if the p-value of below 0.05 in the univariate analysis. All statistical comparisons were two-way, and a p-value of below 0.05 was set as statistical significance.

| Patient selection and characteristics
Initially, 158 ESCC patients treated with PD-1 inhibitors were identified, of which, 35 patients were excluded including 11 patients with one cycle of immunotherapy, 5 patients without treatment assessment, 12 patients without blood results at baseline, 2 patients with stage I, and 5 patients with stage II. Ultimately, 123 patients were used for analysis in this study ( Figure 1). Patient characteristics at baseline are shown in Table 1. Of all patients, the median follow-up was 20.0 months and the median age was 61 years (range: 40-80); 91.1% were male, 82.9% with stage IV, and 66.7% were former smokers; most of the patients (92.7%) with ECOG 0-1. 73.2% received PD-1 inhibitors plus chemotherapy, and 11.4% received radiotherapy; 52 patients (42.3%) received pembrolizumab treatment; first-line treatment, secondline treatment, and third-line and above treatment accounted for 52.8%, 30.9%, and 16.3%, respectively; treatment efficacy of PR, stable disease (SD), and progressive disease (PD) accounted for 39.0%, 41.5%, and 19.5%, respectively; 21 patients (17.1%) had Hb values below 105 g/L, and 40 patients (32.5%) had NLR above 4.3. ECIPI good was in 74 patients (60.2%), intermediate in 37 patients (30.1%), and poor in 12 patients (9.8%), respectively. Until the last follow-up, 43.1% of patients had died.

| Treatment efficacy and survival analysis for ECIPI
As shown in Table 4 Figure 2).

| Subgroup analysis for ECIPI
The differences in characteristics among the three ECIPI groups are shown in Table 5. Variables of ECOG PS, treatment strategy, and treatment line were unbalanced among the three ECIPI groups (p < 0.05). The proportion of ECOG 0-1 (62% vs. 33%, p = 0.035), PD-1 inhibitor plus chemotherapy (67% vs. 42%, p = 0.024), and first-line treatment (80% vs. 37% vs. 40%, p < 0.001) were higher in patients with ECIPI good than that of ECOG ≥ 2, without chemotherapy, second-line treatment, and third-line and above treatment. Subgroup analysis showed that ECIPI good was associated with improved PFS and OS in patients with ECOG 0-1, PD-1 inhibitor plus chemotherapy, firstline treatment, and smoke with all p < 0.05 (Figure 3).

| DISCUSSION
Though EC is one of the most lethal cancers worldwide, its treatment development has made modest improvements in recent years. 17 ESCC accounts for the majority of EC and is a frustrating disease with limited treatment options. The clinical application of anti-PD-1 therapy has revolutionized cancer treatment. PD-1/PD-L1 targeted immunotherapy is considered an innovative treatment strategy with durable anticancer activity and long survival, 18 however, the effectiveness of immunotherapy is below 20%, and most patients cannot experience treatment benefits. 19,20 Although the incidence of immune-related adverse events caused by immunotherapy was low, some were serious and even life-threatening. [21][22][23][24] Therefore, there is a necessity to investigate biomarkers to identify the most suitable patients for such therapy. Biomarkers such as PD-L1 expression and tumor mutational burden were clinically used predictors, but the detection of these biomarkers requires assays and their predictive values are unsatisfactory. 25,26 Convenient and practical biomarker identification is a necessary guarantee for clinical practice.  Anemia is one of the most common hematological abnormalities in cancers, and its incidence increases with the progression of malignancy and the intensification of cancer treatment. 9 Hemoglobin below 110 g/L is considered the diagnostic of cancer-related hemoglobin loss. 27 Previous research reported that anemia in patients with advanced cancer leads to insufficient T-cell responses and induces immunosuppression. 28 In addition, hypoxia induced by Hb reduction may stimulate tumor growth and progression, reduce tumor sensitivity to cancer therapy, and ultimately lead to poor survival. 29,30 Clinical research showed that pretreatment low Hb level was correlated with poor prognosis in cancer patients. [31][32][33] However, several studies reported this association was not significant. 34,35 Currently, no studies have evaluated the predictive role of Hb in ESCC patients with PD-1 inhibitor treatment. Our results suggested that patients with pretreatment low Hb were associated with worse clinical outcomes. Decreased Hb was an independent factor for shortened OS in multivariate analysis, but only a tendency to shorten PFS was observed for the small sample size and interactions of other influencing factors.
Inflammation has a critical value in the occurrence, development, and metastasis of cancer, and increasing evidence link inflammation to the prognosis of patients with cancer. [36][37][38] Elevated neutrophils reflect systemic and local inflammatory responses. Neutrophils provide a favorable microenvironment for tumor cell proliferation and promote the progression and invasion of tumor cells. 11 In contrast, a reduced number of lymphocytes could suppress the immune response to cancer. 39,40 The NLR, calculated based on neutrophils and lymphocytes, has emerged as a potential cancer prognostic biomarker of clinical importance, especially because of its availability and the fact that this relationship can be easily calculated from a patient's routine blood count. Although previous studies investigated the role of NLR, especially its association with the prognosis in various cancers, [41][42][43][44] it has not yet been well evaluated in patients with advanced ESCC with anti-PD-1 therapy. Our findings showed that NLR was an independent influential factor for PFS and OS in multivariate analysis, and patients with pretreatment low NLR could have better PFS and OS than those with high NLR. The ECIPI was developed by combining Hb and NLR. We first evaluated the relationship between pretreatment ECIPI and the prognosis in patients with advanced ESCC receiving PD-1 inhibitors. Our findings indicated that patients with ECIPI good had the best improvements in ORR, PFS, and OS, while those with ECIPI poor had the worst treatment outcomes. Pretreatment ECIPI may be a convenient and useful tool to help identify patients likely to get benefit from PD-1 inhibitor treatment and those who may not obtain treatment benefits.
Some limitations should be considered in this study. First, this study may have some limitations for a lack of clinical data and selection bias. Second, there are other prognostic factors affecting Hb and NLR levels, such as infectious diseases. Our retrospective study failed to stratify infectious diseases, and the impact of NLR and Hb on the prognosis of ESCC patients receiving immunotherapy requires further study. Third, the optimal cut-off values of Hb and NLR are not yet clear, we used X-tile software to calculate the cutoff values based on our data, and the results showed that pretreatment Hb and NLR thresholds of 105 g/L and 4.3 were independently related to treatment outcomes. Finally, although some characteristics, such as ECOG PS, treatment strategy, and treatment line, may influence the final results, we performed a subgroup analysis and the results were still significant. Nevertheless, this study provides a useful and convenient tool to identify patients likely to respond to PD-1 inhibitors and those who may be resistant to these drugs. Further research is warranted to clarify these findings. Third-line and above 8 6 6 Abbreviations: ECIPI, esophageal cancer immune prognostic index; ECOG PS, Eastern Cooperative Oncology Group performance status; ESCC, esophageal squamous cell carcinoma.
T A B L E 5 Comparisons of characteristics among the three ECIPI groups.

| CONCLUSION
Pretreatment ECIPI was correlated with the prognosis in advanced ESCC patients treated with PD-1 inhibitors, suggesting that ECIPI is a useful and convenient tool for identifying patients who may benefit from PD-1 inhibitors.